ZodiacMSM: A Heterogeneous, Multi-node and Scalable Multi-Scalar Multiplication System for Zero Knowledge Proof Acceleration

被引:0
作者
Xu, Yiyang [1 ]
Qian, Dahong [1 ]
机构
[1] ZKFlash Labs Inc, San Jose, CA 95112 USA
来源
2023 IEEE 36TH INTERNATIONAL SYSTEM-ON-CHIP CONFERENCE, SOCC | 2023年
关键词
Zero-Knowledge Proof; zk-SNARK; ASIC; MultiScalar Multiplication; Heterogeneous System; scalable acceleration; ZKP;
D O I
10.1109/SOCC58585.2023.10257155
中图分类号
TP301 [理论、方法];
学科分类号
081202 ;
摘要
In cryptography, Zero-knowledge proof (ZKP) is a protocol that enables a party to demonstrate the accuracy of a given statement without revealing any additional information beyond the validity of the statement itself. This protocol ensures both the confidentiality and integrity of the computation, and has therefore become widely utilized in various industries for privacy-preserving, including online age verification and electronic voting. The most computationally demanding component of most Zero Knowledge Proof systems is the Multi-Scalar Multiplication (MSM) module. To address this issue and achieve low power and cost efficient MSM for different applications, a novel heterogeneous MSM architecture, ZodiacMSM, is developed. The architecture utilizes the Pippenger algorithm and a multi-chip design with memory access partitioning, which enables scalable performance for polynomials of various sizes in different ZKP applications. The performance of this architecture increases almost linearly with the number of nodes. ZodiacMSM supports multiple elliptic curves through reconfigurable fully-pipelined point-addition to achieve lower latency and minimize scheduler effort. When evaluated in 28nm, ZodiacMSM is able to achieve 2.3x speedup on MSM with single node and 32x speedup on 16 nodes compared to previous peer-reviewed state-of-the-art ASIC results [8].
引用
收藏
页码:330 / 335
页数:6
相关论文
共 13 条
  • [1] Aasaraai Kaveh, 2022, Paper 2022/1396
  • [2] Zero-Knowledge Contingent Payments Revisited: Attacks and Payments for Services
    Campanelli, Matteo
    Gennaro, Rosario
    Goldfeder, Steven
    Nizzardo, Luca
    [J]. CCS'17: PROCEEDINGS OF THE 2017 ACM SIGSAC CONFERENCE ON COMPUTER AND COMMUNICATIONS SECURITY, 2017, : 229 - 243
  • [3] Gabizon A., 2019, Plonk: Permutations over lagrange-bases for oecumenical noninteractive arguments of knowledge
  • [4] github, bellperson
  • [5] THE KNOWLEDGE COMPLEXITY OF INTERACTIVE PROOF SYSTEMS
    GOLDWASSER, S
    MICALI, S
    RACKOFF, C
    [J]. SIAM JOURNAL ON COMPUTING, 1989, 18 (01) : 186 - 208
  • [6] Groth J, 2005, LECT NOTES COMPUT SC, V3531, P467
  • [7] On the Size of Pairing-Based Non-interactive Arguments
    Groth, Jens
    [J]. ADVANCES IN CRYPTOLOGY - EUROCRYPT 2016, PT II, 2016, 9666 : 305 - 326
  • [8] zkCNN: Zero Knowledge Proofs for Convolutional Neural Network Predictions and Accuracy
    Liu, Tianyi
    Xie, Xiang
    Zhang, Yupeng
    [J]. CCS '21: PROCEEDINGS OF THE 2021 ACM SIGSAC CONFERENCE ON COMPUTER AND COMMUNICATIONS SECURITY, 2021, : 2968 - 2985
  • [9] minerdaily, Does the Antminer S19 XP use 5nm ASICs
  • [10] Pippenger N., 1976, 17th_Annual IEEE_Symposium_on_Foundations_of_Computer_Science, P258